“Change is what can actually make the tough times better”: A patient‐centred patient safety intervention delivered in collaboration with hospital volunteers

Abstract Background The PRASE (Patient Reporting and Action for a Safe Environment) intervention provides a way to systematically collect patient feedback to support service improvement. To provide a sustainable mechanism for the PRASE intervention, a 2‐year improvement project explored the potential for hospital volunteers to facilitate the collection of PRASE feedback. Objective To explore the implementation of the PRASE intervention delivered in collaboration with hospital volunteers from the perspectives of key stakeholders. Design A qualitative case study design was utilized across three acute NHS trusts in the United Kingdom between March 2016 and October 2016. Ward level data (staff interviews; action planning meeting recordings; implementation fidelity information) were analysed taking a pen portrait approach. We also carried out focus groups with hospital volunteers and interviews with voluntary services/patient experience staff, which were analysed thematically. Results Whilst most ward staff reported feeling engaged with the intervention, there were discordant views on its use and usefulness. The hospital volunteers were positive about their involvement, and on some wards, worked with staff to produce actions to improve services. The voluntary services/patient experience staff participants emphasised the need for PRASE to sit within an organisations’ wider governance structure. Conclusion From the perspective of key stakeholders, hospital volunteers facilitating the collection of PRASE feedback is a feasible means of implementing the PRASE intervention. However, the variability around ward staff being able to use the feedback to make changes to services demonstrates that it is this latter part of the PRASE intervention cycle that is more problematic.

intervention was developed and tested between 2010 and 2015.
In order to provide a sustainable mechanism for the PRASE intervention beyond a research study context-a 2-year improvement project subsequently explored the potential for hospital volunteers to facilitate the collection of feedback within PRASE. The PRASE in collaboration with hospital volunteers project was implemented at three NHS trusts (at three hospital sites), across selected wards.
A two-stage evaluation ran alongside implementation. The results of the formative evaluation which explored the perceptions of key stakeholders throughout the pilot phase have been published previously. 16 A summative evaluation followed, which coincided with the PRASE intervention being scaled up and spread across additional wards. This evaluation phase focussed on (a) the ward level experience and (b) the ongoing experiences of hospital volunteers and voluntary services/patient experience staff. We aimed to explore the implementation of PRASE in collaboration with hospital volunteers from the perspectives of these key stakeholders, to gain an in-depth understanding of each ward's PRASE journey, and a collective account of implementation.

| Research questions
• What is the ward level experience of the PRASE intervention delivered in collaboration with hospital volunteers?
• What are the on-going experiences of hospital volunteers, voluntary services and patient experience staff involved in the PRASE intervention?

| Patient involvement
A full description of patient involvement in the design and conduct for the wider project has been published previously. 16 The research aim initially arose from discussions about the sustainability of the PRASE intervention with patients and healthcare professionals.
Additionally, a patient representative was part of the initial application for funding, and was invited to attend all steering group meetings.

| Design
Given the relatively small number of participating wards (N = 7), we utilized a case study approach. 17 This combined different types of data including the following: semi-structured interviews with ward staff (one-one and dyad interviews); researcher notes of APM audio recordings; and information pertaining to implementation fidelity. For each ward, we aimed to produce a synthesised account of their PRASE journey. In order to generate a means of assessing implementation fidelity, the key PRASE ward activities were predefined by the implementation team (see Figure 1). These activities were based on a version of the PRASE intervention programme theory agreed by the wider project team at the end of the pilot phase. 16 Focus groups with hospital volunteers and semi-structured interviews with voluntary services/patient experience staff were conducted at two time points and focussed on the participants' experiences of their involvement in PRASE and their views on implementation, with specific questions varying for each participant group.

| Setting and sample
Three acute NHS trusts in the United Kingdom were involved in the improvement project, and data collection took place between March 2016 and October 2016. The project was led by a central project team at the lead trust (trust 1). Seven wards were involved in the summative evaluation (roll-out wards). We refer to each of the wards by pseudonyms (Ward A -Ward G). APMs were digitally recorded for five of the seven wards; two wards did not complete an APM, and eight staff (from four wards) participated in interviews (range 9-39 minutes; average 26 minutes). We conducted four focus groups with hospital volunteers (n = 13), A and seven interviews with voluntary services/patient experience staff at the beginning and end of the roll-out phase (n = 5). B More information regarding the data sources, characteristics of ward and study participants and the characteristics of the NHS trusts and services/departments involved in delivering the project are presented in Table 1.

| Evaluation procedure
Hospital volunteer participants were invited to take part in the focus groups by a member of the local project implementation team. Staff participants (ward, voluntary services and patient experience) were approached directly by an evaluation researcher and invited to participate. Participant information sheets were distributed in advance, and written informed consent obtained from all participants. Focus groups and interviews took place in a private room on the hospital sites. The implementation team facilitator recorded ward APMs. C At the start of APMs, the implementation team facilitator confirmed with the participants that they were happy for the meeting to be audio recorded for evaluation purposes.

| Data analysis
In an earlier process evaluation of the PRASE intervention, Sheard and colleagues 15 report how pen portraits have been used in applied health research in qualitative studies to provide a narrative account of a typical participant, or as an analytic aide memoir.
Recognising the lack of methodological literature around the construction of a pen portrait, they describe how they "created a basic structure for the pen portraits which centred on the writing of a linear, longitudinal account of how each ward had engaged with relevant key components of the intervention and the contextual factors which influenced this." 15(p3) This previous work guided our approach, and we endeavoured to produce a rich account of the journey of each ward.
The ward staff interviews were digitally recorded and transcribed verbatim, and three researchers listened to the APM recordings and made detailed notes (JOH, LH and GL), with a particular emphasis on the role of hospital volunteers in the ward APMs. An account of implementation fidelity was also produced (see Table 2).
We developed a proforma to facilitate the synthesis of these data sources. The proforma aimed to elicit specific information for each ward, for example, timeline of PRASE activities, staff views regarding: hospital volunteer involvement; APMs (eg how actions were decided, role of facilitation); ward involvement in other initiatives.
Although the proforma specified a priori information of interest, the approach also allowed for emergent concepts. Three researchers (JOH, LH and GL) synthesised these data sources to produce a pen portrait narrative account of each ward's PRASE journey. A broader level synthesis of the ward pen portraits was generated by two researchers (JOH and GL). An example pen portrait for Ward A is provided in Appendix 1.
The hospital volunteer focus groups and voluntary services/patient experience staff interviews were digitally recorded and transcribed verbatim. In a series of analysis meetings, two researchers (JOH and GL) took a thematic approach to analysis allowing for both a priori and emergent concepts and themes, with disagreements resolved through discussion.
Subsequently, the same two researchers met in an intense analysis session to generate meta-themes. The aim of this session was to have a discussion of commonality and differences across the data at a meta, abstract level in order to synthesise the findings from the broader level synthesis of the ward pen portraits and the themes from the hospital volunteer focus groups and voluntary services/patient experience staff interviews.
A Several hospital volunteers participated in more than one focus group.

| Ethics and governance
The appropriate governance approvals were sought for each research site, and ethical approval was granted by the University of Bradford, Humanities, Social and Health Sciences Research Ethics Panel. D

| PRASE intervention
The development and testing of the PRASE intervention has been fully described in previous published work. 7,[9][10][11][12][13][14][15] We present a brief description below:   b. lagging indicators of safety (PIRT 7,9 ) (ii) patient feedback collated into a feedback report (iii) report considered within a multidisciplinary action planning team. E (iv) action plans made The intervention was designed to be cyclical, and time frames are not specified for these activities above what is described in the published literature, which reports 6-month cycles. F

| RE SULTS
We now describe the five meta-level themes. Not all wards progressed through the planned implementation of the intervention and participation in evaluation activities differed at trust and ward level. For instance, volunteer focus groups and voluntary services/ patient experience staff interviews were not held at trust 3, and APMs did not take place within the summative evaluation phase at this site. Therefore, themes 1-4 represent a synthesis of the available data sources, which principally relate to trusts 1 and 2, and theme 5 relates to all trusts.  TA B L E 2 Implementation fidelity in relation to key PRASE ward activities Activity Implementation fidelity

| Legitimacy and validity of PRASE
(1) Key stakeholders attend project start-up meeting prior to commencement of roll-out The agreed multidisciplinary action planning team receive the feedback report and supporting guidance/documentation to facilitate the APM

| Challenges of using patient feedback to support service improvement
Ward staff often focussed their efforts on identifying smaller changes that were within their influence, and easier for them to shape the outcome. Such quick fixes were more frequently agreed than systemic solutions. This may have been for a number of reasons. Although perceptions of APM facilitation were often positive, at times, facilitation seemed to encourage a focus on every domain within the report. This may have inadvertently led to a higher volume of actions, at the expense of a more nuanced or systemic approach to change. It is also possible that the lack of multiprofessional approach (grades and disciplines) within some wards perhaps led to a narrower focus or range of solutions. There was an articulated sense of frustration from some staff that PRASE in and of itself cannot supply the resources and support needed to enact real change, and that the intervention, like many improvement efforts, ended up going by the wayside during periods of strain or uncertainty.
There was a general feeling that complex change requiring increased finances or resources were not well supported within the structure of PRASE. Some areas for action were reported as being 'out of their hands', this was particularly apparent if other teams and services were required to be engaged, or the action required increased resources such as staffing. One clinician summed up their involvement in PRASE stating: It's not easy and when your back is against the wall….
things like these change projects are the things that

stop….but it means that change doesn't occur and change is what can actually make the tough times better…
Ward A pen portrait excerpt

| D ISCUSS I ON
This study explored the experiences of ward staff, hospital volunteers, and those supporting them, in the implementation of a patientcentred patient safety intervention in collaboration with volunteers.
The PRASE intervention had credibility with ward staff, although this was not universal. The legitimacy of patient feedback for service improvement was, however, unanimous across all participants.
The role of the hospital volunteers in this intervention was valued by all stakeholders, principally for their independence, and the transparency and accountability that arose through their involvement in APMs. The importance of targeted volunteer recruitment and on-going support, including feedback on performance, were reinforced. Volunteers were keen to close the loop by attending APMs.
We found significant challenges identified in terms of planning and implementing service improvements as part of the PRASE cycle.
Finally, we noted differences in the nature of the implementation of the intervention across the three study sites, which may have impacted on the relative success of the intervention. These findings raise a number of interesting issues that we will now discuss further.
The findings relating to the infrastructure, training and support mechanisms required for the implementation of PRASE with hospital volunteers were consistent with the wider literature on the importance of understanding volunteers' motivations and meeting their expectations for retention. 18,19 Many of these findings reinforced the key themes from the formative evaluation, 16 and therefore, the discussion reflects more so on novel insights.
Whilst the intervention was received positively by most staff, others were less convinced by its utility and value in supporting safety improvements. It is difficult to decouple this variation from the different implementation approaches. However, we can perhaps be more certain that where PRASE worked best was where relations were good with the volunteers, the feedback was timely, and staff met as a multidisciplinary team to consider the feedback from patients and attempted to make changes using the feedback. Problems arose when these data were not timely, or when there was too much data. The nature of the implementation does seem to impact on the experience of staff involved. The differences between wards in terms of progress through the implementation activities may be due to the implementation team sitting within the lead trust, with implementation activity therefore having a somewhat diluted effectiveness outside of this trust. Further evidence of the key role of implementation is in the need for facilitation of APMs, which mirrors the early developmental work. 9 Whilst facilitation was regarded positively by staff at the lead trust, there was a less positive perception elsewhere, which perhaps may have been interpreted as undermining the local ownership of the project. These findings speak to issues of complexity, which have been described as "resulting from interactions among many component parts-is a property of both the intervention and the context (or system) into which it is placed." 20 (p307) There is recognition that defining the key components of an intervention with a view to standardize across sites may not always be the best approach. Indeed, some authors propose standardizing by function, meaning there may be variation in the form components take across sites, but the function a component performs in the local context should be consistent. 21 Therefore, in our work, it is possible that implementation fidelity may not be an accurate indication that the function of the activities was achieved.
There is increasing attention on how staff use and act on data within health services. 6,14,15,22 We found an unequivocal lack of movement from data to action within our case study wards, the reasons for which are likely to be manifold and complex. A significant challenge for the implementation of this intervention seemed to be getting health professionals together to discuss, interpret and act on safety-related data. Such a finding strongly resonates with previous published work describing the PRASE intervention. 9,14,15 The need to create space and competence for improvement by healthcare staff has been argued as crucial if health services are to improve care and increase efficiency. 23 With respect to patient feedback specifically, a key challenge for engaging patients in safety or service improvement is using the data as a basis for meaningful change. 24,25 Indeed, some authors have reported this challenge as a chasm between the activity related to collecting patient feedback, and the complete lack of such feedback filtering down to, and being used by, frontline clinical staff. 26  Other conditions for effective A&F interventions are thought to be targeting behaviours that are easy to change, and that the data do not overwhelm and are as simple and specific as possible. Given the almost limitless range of issues identifiable from the combination of PMOS scores and PIRT safety concerns, it is clear that the PRASE feedback report is unlikely to be regarded as simple data, with required changes necessarily within the gift of ward staff. Facilitating the APMs was designed to reduce the potential cognitive load and support directed effort, 9 but it would appear from our findings that this has not had the desired effect. However, the facilitation role may have been important for a further hypothesis-that A&F interven- One further issue of relevance to our findings is the nature of the PRASE data, and whether it-as soft intelligence-fits within the current dominant improvement approaches within hospitals.
Soft intelligence has been described as "the processes and behaviours associated with seeking and interpreting soft data-of the kind that evade easy capture, straightforward classification and simple quantification-to produce forms of knowledge that can provide the basis for intervention." 6(p19) It has been argued that seeking out data which does not easily conform to measurement is a necessary part of managing patient safety, 28 even when it is "discomfiting and disruptive." 6(p26) Our findings seem to resonate with previous suggestions that patient feedback is soft data. 6,22 PRASE feedback is likely to generate uncertainty rather than certainty in terms of the problems to be solved, requiring further interrogation of extant data or other data gathering. Indeed, within our study, at an organisational level, questions were raised in terms of where PRASE sits within an organisations' wider governance structure, and the services and departments involved in co-ordinating its delivery, both of which may facilitate PRASE with volunteers being successfully embedded within an organisation. These issues, combined with the sense of unease patient feedback about their experience of safety may create, perhaps make it less surprising that staff were unable to make anything more than quick fixes based on the report.

| IMPLIC ATIONS
We present some general recommendations that may support healthcare organisations seeking to implement the PRASE intervention with hospital volunteers: (1). establish a core team locally that has ownership, to ensure implementation is sensitive to the local context to avoid confusion and reduce dilution; (2). ensure the timeliness of patient feedback to ward staff, to increase the legitimacy of the data and support engagement with subsequent action planning;

| LI M ITATI O N S
The main limitation of this work is that we had limited ability to draw conclusions regarding implementation at trust 3. We were unable to hold volunteer focus groups, with factors such as project progress, volunteer availability and attrition contributing to this. Wards at this site did not hold an APM, and thus, we were unable to interview ward staff. As we could only include in our research the organisations involved in the wider project, this meant we were limited in terms of the number of wards involved, and consequently the number of data sources we could draw upon.

| CON CLUS IONS
Our findings suggest that from the perspectives of key stakeholders, hospital volunteers as conduits of patient feedback about safety are a feasible means of implementing the PRASE intervention. However, the variability and complexity we found in the ability of ward staff to use the feedback to make changes to services demonstrates that it is this latter part of the PRASE intervention cycle that is more problematic.

AC TI O N PL A N N I N G
The APM was attended by senior nurses, medics and volunteers, although ideally APM participants highlighted that they would have chosen to also involve more junior staff. Volunteers had a strong voice at the APM and staff commented that they valued their clinical independence and ability to contextualise information in the report.
The facilitator took the meeting members through the detail of the action plan domain by domain. Nine problems emerged, and 21 individual points to action with 18 of the action points falling to the senior nurses to action. In general, discussion concentrated on negative comments rather than the report as a whole. The majority of actions were simple to implement, but finding spare capacity to fulfil complex actions was more difficult. Focus moved towards the difficulties of improving issues, and there was also concern regarding the need for longer time frames to embed changes due to the staffing shortages:

It needs to go at their pace because what I don't want
them to feel is that there is another thing that's being put onto them when they're already struggling. So, you know, if it takes us longer to make that change than it would normally do then we'll just have to go at that pace.
Staff appeared to be confident in identifying more immediate "quick fix" actions over which they felt they had influence. Such actions included ordering badges and decaffeinated teabags, although even these were thought to be potentially problematic to implement.
The ward was positive about points for action that already had a solution; for example, "poor response to buzzer" was an area for action and the ward already had a new buzzer system in place.

BA R R I E R S A N D FACI LITATO R S
The ward appreciated that a named, skilled, driven facilitator was needed to act as a catalyst for the PRASE process to maintain momentum. After the initial process is established, they felt that the role could be filled by a suitably skilled member of the ward team. There was a general feeling that complex change requiring increased finances or resources were not well supported within the structure of PRASE.
Some areas for action were reported as being "out of their hands," this was particularly apparent if other teams and services were required to be engaged, or the action required increased resources such as staffing. One clinician summed up their involvement in PRASE stating: It's not easy and when your back is against the wall….
things like these change projects are the things that

stop….but it means that change doesn't occur and change
is what can actually make the tough times better…